Cancer Diagnosis and Therapy
Health profiles in 5836 long-term cancer survivors
Version of Record online: 23 JAN 2003
Copyright © 2003 Wiley-Liss, Inc.
International Journal of Cancer
Volume 104, Issue 4, pages 488–495, 20 April 2003
How to Cite
Schultz, P. N., Beck, M. L., Stava, C. and Vassilopoulou-Sellin, R. (2003), Health profiles in 5836 long-term cancer survivors. Int. J. Cancer, 104: 488–495. doi: 10.1002/ijc.10981
- Issue online: 7 FEB 2003
- Version of Record online: 23 JAN 2003
- Manuscript Accepted: 23 NOV 2002
- Manuscript Revised: 20 NOV 2002
- Manuscript Received: 23 MAY 2002
- adult cancer;
- late effects;
Increasingly, prolonged survival follows the diagnosis of cancer. Cancer therapies result in complex and lasting health effects that create unique health-care needs for the survivors but are poorly understood (especially in survivors of adult cancers). Cancer survivors were asked to respond to a mailed health survey and provide medical and social information pertaining to their cancer experience. Information about demographics and perceived disease-related medical problems was analyzed. We analyzed the response of 5,836 survivors of adult cancers. Two-thirds of the responses came from women, and the response rate was 51% in both sexes. The mean interval since cancer diagnosis was 18.0 ± 8.5 years. Younger survivors and men were more likely to report that cancer had affected their health. The health effect most commonly reported by survivors was arthritis/osteoporosis (26% of respondents). Survivors of Hodgkin's disease prominently reported thyroid and lung problems (33.8% of responders with the diagnosis). Prior diagnosis of lymphoma was associated with frequent mention of memory loss (14.7%). The passage of time decreased some perceived effects (memory loss) but increased others (arthritis/osteoporosis, cataracts). Compared with the general population, the incidence of several age- and gender-adjusted health conditions in cancer survivors is different. This group of cancer survivors reported generally good health but outlined multiple lasting medical problems. The health survey described represents 1 approach to the development of comprehensive information about the health needs of cancer survivors. © 2003 Wiley-Liss, Inc.
More people than ever are surviving cancer and are living longer thanks to improvements in early detection, treatment and general supportive care. Indeed, cancer deaths have declined in both men and women since 1992.1 The National Cancer Institute estimates that approximately 1.3 million patients are diagnosed with invasive cancer every year (16 million new cases since 1990), with an overall 5-year survival rate of 62%.2 In fact, 8.9 million Americans with a history of cancer were alive in 1997. As their numbers and long-term health care needs increase, these cancer survivors are expected to require an increasingly large portion of public health resources.
For example, cancer is quite rare in children, with fewer than 10,000 new cases per year, yet 5-year survival rates have been increasing and now range between 70% and 92%.2 It is estimated that 1 in 1,000 Americans between the ages of 15 and 45 was a survivor of childhood cancer in 1990; this rate was anticipated to become 1 in 900 persons in 2000 and possibly as high as 1 in 250 persons in 2010.3 In Britain, approximately 60% of children with cancer are cured; as a result, it is estimated that 1 in 1,000 Britons in the general population represent survivors of childhood cancer.4 This population will reach adulthood with a near-normal life expectancy and with a background of complex and often intensive antineoplastic therapy. More than 5 million Americans have been treated for cancer and are alive for at least 5 years since the diagnosis.
Attention is increasingly directed toward understanding the long-term impact of cancer and cancer therapy, especially as it may affect the quality of life of cancer survivors. Literature addressing psychosocial late effects is gradually becoming robust for survivors of childhood and adult cancers. Information regarding the lasting medical late effects of cancer and cancer therapy, however, remains very limited.
Several investigators have highlighted the presence and complexity of lasting, medical late effects throughout the adult lives of childhood cancer survivors.5, 6, 7 Stevens et al.8 reported that 58% of survivors had at least 1 chronic health problem, and 32% had 2 or more, including second primary cancers. Much less is known about the lasting medical impact of cancer treatment in survivors of adult cancers. Investigators addressing physiologic late effects of cancer9, 10 highlight the paucity of information, while emphasizing that multiple systems are often affected. For example, pelvic surgery can impair fertility, whereas splenectomy increases susceptibility to bacterial infections. Certain chemotherapy agents and regimens may compromise the health of the cardiac, pulmonary, genitourinary, endocrine and neurologic systems. Radiotherapy may potentiate these complications and impair skeletal development and dental health in addition to contributing to the risk of second primary cancers. Combining treatments can increase the number or severity of medical late effects.
Most reports concerning the long-term health effects of adult-onset cancer patients cover an interval of less than 5 years from diagnosis.11, 12 For example, 1 study described the sexual, bowel and urinary function in prostate cancer patients after radiation up to 24 months post treatment.13 Another study followed 277 testicular cancer survivors from 3 to 13 years and assessed for backache, leg pain, cough, eye problems, infertility and self-esteem.14 Sarna et al.15 reported comorbidity in >60% of 142 lung cancer survivors. Another study reported that 20% of ovarian cancer survivors had abdominal and gynecologic side effects and neurotoxicities.16 Others have reported on late effects in long-term survivors of Hodgkin's disease including children.17, 18, 19, 20 Overall, the available studies tend to focus on relatively specific health questions, to pertain to individual cancer types and to have fairly brief follow-up duration.
Hence, it is critical to develop a comprehensive body of knowledge regarding the health profiles of different groups of cancer survivors so that appropriate strategies can be developed for their health care. The present report describes the approach we have adopted at The University of Texas M. D. Anderson Cancer Center, with which we hope to begin defining any lasting effects of cancer treatment on the health and welfare of cancer survivors. The material presented here focuses on medical health effects reported by survivors of adult cancers.
MATERIAL AND METHODS
At the M. D. Anderson Cancer Center, we have developed the Life After Cancer Care (LACC) program, which focuses on the cancer and cancer treatment-related health concerns of cancer survivors. Within the context of the LACC program, we developed a survey with which to gather information about the health profiles of cancer survivors. We sought information from individuals who had a history of cancer diagnosed at adult age, who had completed all antineoplastic treatments, who had survived at least 5 years after the diagnosis of cancer and who were still living without evidence of malignant disease.
Two sources of patient participation were identified. The first consisted of former M. D. Anderson patients who were known to be alive more than 5 years after diagnosis and who were no longer receiving active treatment at M. D. Anderson; they were identified through a search conducted by the Department of Medical Informatics. A survey was mailed to the last known addresses of these patients, with a self-addressed stamped envelope enclosed for its return. In addition, the survey was posted on the LACC Web site http://www.mdanderson.org/departments/lacc/, and individuals had the opportunity to participate by responding through that electronic medium. Computer information technology was utilized to minimize duplicate responses; the primary keys for the database were the M. D. Anderson medical record number (mailed surveys) and the registration number plus date of survey completion, which facilitated removal of multiple entries. Our Institutional Review Board approved this research proposal, including the survey.
Description of the survey
The survey asked for demographic data such as date of birth, sex, marital status and ethnic group. Specific questions were included about the survivors' cancer history, such as date of diagnosis, cancer type and cancer treatment(s). The types of cancer from which survivors were asked to choose included breast cancer, gynecologic cancer, melanoma, genitourinary cancer (including prostate cancer), head and neck cancer, colorectal cancer, thyroid cancer, non-Hodgkin's and Hodgkin's lymphoma, soft tissue and bone sarcoma, lung cancer, acute and chronic leukemia, gastrointestinal cancer, unknown primary cancer, central nervous system cancer, neuroendocrine cancer, multiple myeloma and other. Cancer treatment options included surgery, chemotherapy, radiotherapy and other. The survey was developed from an intake assessment tool used by patients attending the outpatient clinics at M.D. Anderson Cancer Center.
The survey included questions about medical health effects. A general question was, “Do you believe that having cancer has affected your overall health?” A checklist of specific health conditions and/or effects was included and participants were asked to check all that applied. These included abdominal pain, arthritis/osteoporosis, bleeding problems, cataracts, circulation problems, diabetes or sugar in the urine, dizziness, frequent infections, hearing loss, heart problems, HIV/AIDS, kidney/urine problems, liver problems, loss of memory, lung problems, migraine headaches, psychological or psychiatric problems, seizure, skin disorder, stroke, thyroid problems, synchronous cancers and other. At the end of this list was a blank space for comments.
Sociocultural questions were also included in the survey. Several questions inquired about the survivors' family dynamics. Respondents were asked if having cancer had improved or hurt their family and intimate relationships. Survivors were asked about job- and work-related issues, education, insurance and medical care. These psychosocial items are not included in the present analysis and will be presented separately. Surveys were addressed to survivors of adult cancers (present report) and to survivors of childhood cancer, or their parents (also to be presented elsewhere). The data were entered into an Oracle Enterprise Edition Version 8.1.5, which is maintained and updated on a regular basis (by the institution's information team and by C.S.). The present report focuses specifically on information pertaining to the medical health profiles as reported by survivors of adult cancers.
STATISTICA for Windows Version 6.0 by StatSoft (Tulsa, OK) was used for statistical analyses. Frequency distributions and histograms were used to describe the distributions of the responses. Descriptive statistics, including percentages, mean, standard deviations and ranges were used to summarize the information gathered from the survey. Chi-square tests, correlational associations and univariate analyses were used as appropriate. Indices of reliability including the Cronbach index were computed for the scales contained in the survey. The National Health Interview Survey (NHIS), 1997, is a multistage probability sample survey conducted annually by interviewers of the U.S. Census Bureau for the National Center for Health Statistics, Centers for Disease Control and Prevention and is representative of the civilian noninstitutionalized U.S. population.21 The NHIS was used to compare the prevalence of chronic health conditions in the U.S. population as a whole with their prevalence in cancer survivors in our study. The prevalence of specific chronic conditions of cancer survivors from our study were converted into individuals/1,000 and compared with the NHIS prevalence in the U.S. population. An estimate of relative risk was calculated to determine if cancer survivors had a higher relative risk of specific chronic conditions identified by the survey.
Surveys were mailed to survivors of adult cancer (≥18 years of age at diagnosis) who had, at some point, registered at M. D. Anderson and who met the previously described criteria. During the initial mailing, 4,830 surveys (50.8%) were returned within 15 months of mailing. Additional mailings of surveys were done to M.D. Anderson patients as they reached the 5-year survivor criteria. The total number of mailed surveys returned is 5,935 survivors of adult-onset cancer. There was no mechanism in place to determine the fate of those surveys that were not returned. Survivors may have moved, changed their name or chosen not to respond. During the same period, survivors of adult-onset cancers completed 1,514 surveys electronically via the web site. This provided data from 7,449 cancer survivors. Surveys with incomplete demographic data or web entries that did not meet the inclusion criteria were eliminated from this analysis. We therefore had information derived from a total of 6,388 survivors of adult cancers.
Comparing mailed survey responses and web-based survey responses
Significant differences were noted between web-based surveys (552 entries) and the mailed surveys (5,836 entries). More male cancer survivors responded by mail than via the Internet (37.1% vs. 27.2%, p < 0.00001). Survivors responding to the mailed surveys were significantly older at the time of cancer diagnosis (47.7 ± 14.0 years vs. 38.0 + 12.3 years, p < 0.00001), and they had a longer interval since diagnosis (18.0 ± 8.5 years vs. 11.6 ± 7.2 years, p < 0.00001). Individuals responding via the Internet were more frequently white than African-American (p < 0.00001). There were also differences in the types of cancer between the 2 groups. For example, individuals who responded to the mailed surveys included more survivors of genitourinary cancer (14.8% vs. 6.5%), gynecologic cancer (16.6% vs. 5.4%), head and neck cancer (6.4% vs. 2.2%) and melanoma (13.2% vs. 4.4%). The sample from the web-based surveys included more survivors of Hodgkin's disease (7.3% vs. 2.7%) and thyroid cancer (27.4% vs. 2.6%, p < 0.00001). In addition, more cancer survivors responding via the Internet indicated that having cancer had affected their overall health (70% vs. 32%, p < 0.00001) than those who responded by mail. Because of these significant differences between web-based and mailed surveys, we elected to exclude the web-based responses from this initial descriptive analysis and to focus on the information obtained from the mailed surveys (5836 cancer survivors).
The overall sample included surveys from 3,674 women (62.9%) and 2,162 men (37.1%). The response rates for the mailed surveys were 50.8% for women and 50.9% for men. There were 5,360 patients (91.8%) who classified themselves as white, 180 African Americans (3.1%), 244 Latino/Hispanic (4.2%), 31 Asian Americans (0.5%), and 21 American Indians (0.4%). Their mean age at the time of cancer diagnosis was 47.7 ± 14.0 years. Figure 1 shows the distribution of the age at diagnosis.
The interval between a diagnosis of cancer and completion of the survey was at least 5 years (by design) and generally much longer, with a mean interval of 18.0 ± 8.5 years; Figure 2 shows that distribution. Cancer had been diagnosed more than 25 years earlier in 20% of the patients who filled out the survey, providing information about very long-term effects. Cancer diagnoses were divided into 19 categories, which are outlined in Table I. The distribution of diagnosis reflects in part the likelihood of long-term survival with different cancers, referral patterns to the Cancer Center and selection bias to survey response. Breast cancer was the most common diagnosis (22.3% of all cases) followed by other gynecologic cancers (16.6%), explaining, in part, why more surveys were received from women than from men. Overall, 4 diagnoses (gynecologic cancer, genitourinary cancer, breast cancer and melanoma) constituted 67% of the sample.
|Cancer type||No. of patients||%|
|Head and neck||372||6.4|
|Bone/soft tissue sarcoma||259||4.4|
|Central nervous system||25||0.4|
Survey participants were asked, “Do you believe that cancer has affected your overall health?” Of the 5,709 participants who answered this question, 1,814 (31.1%) responded affirmatively; they were younger at the time of cancer diagnosis (46.8 ± 13.7 years vs. 47.9 ± 14.1 years, p = 0.006), and more frequently male (34% of men vs. 30% of women, p = 0.004). Survivors who answered “no” had a longer mean interval since the time of diagnosis (18.3 ± 8.5 years vs. 17.2 ± 8.3 years); no correlation could be found between time from diagnosis and self-reports of health effects (r = 0.08, p = 0.7) even though the highest incidence of health effects occurred in those survivors who were diagnosed with cancer 5–10 years earlier.
Univariate analysis of cancer type and response to the “did cancer affect your health” question indicated that response was influenced by cancer type (F [19, 5689] = 16.5, p < 0.00001). Among survivors who indicated that cancer had affected their health, the most frequent diagnoses were breast cancer (21.1% of affirmative responses), genitourinary cancer (16.6%) and gynecologic cancer (14.7%). Among survivors who indicated that cancer did not affect their health, the most frequent diagnoses were also breast cancer (22.7% of negative responses) and gynecologicl cancer (17.5%), followed by melanoma (16.9%). This distribution probably reflects the frequency and the natural history of the different cancer diagnoses (survivors of common and/or chronic cancers may appear overrepresented in the sample).
To control for the above, we also looked at the responses relative to the particular cancer types. Survivors were less likely to report that cancer had affected their health if they were previously treated for melanoma (86.6% of melanoma survivors), gynecologic cancer (71.9%) or colorectal cancer (72.9%). Survivors were more likely to report that cancer had affected their health if they were previously treated for multiple myeloma (76.2% of myeloma survivors), central nervous system (62.5%) or gastrointestinal cancer (58.6%); these diagnoses represent a small portion of the sample and results should be interpreted with caution. An almost even distribution was reported from survivors of chronic leukemia, lung cancer and Hodgkin's disease. Patients with a history of melanoma were the least likely to report that the disease affected their health, and patients with multiple myeloma were the most likely. These results are summarized in Table II.
|No. of patients||1,814||3,895|
|Age at diagnosis (yr)||46.8 ± 13.7||47.9 ± 14.1||0.006|
|Time from diagnosis (mo)||17.2 ± 8.3||18.3 ± 8.5||<0.00001|
|Women||1,091 (30.0)||2,496 (70.0)||0.004|
|Men||723 (34.0)||1,399 (66.0)|
|Cancer type||No. of pts.||% all yes responses||% of cancer type||No. of pts.||% all no responses||% of cancer types|
|Head and neck||129||7.1||35.6||233||6.0||64.4|
|Central nervous system||16||0.9||62.5||9||0.2||37.5|
With respect to specific health effects/problems, the most frequently reported was arthritis/osteoporosis (1541 survivors, 26.4%). Table III illustrates the distribution of these self-reported health effects. Gender was an important variable in the survivors' responses regarding their overall health; specific differences also became apparent when specific health effects were analyzed. Men generally reported fewer specific problems; the most frequently mentioned were heart problems (17.2% of responding men), hearing loss (14.9%) and arthritis/osteoporosis (15.5%). Women generally reported more specific problems; the most frequently mentioned were arthritis/osteoporosis (32.9% of responding women), heart problems (13.9%) and thyroid problems (12.1%). There was a gender difference in the likelihood of reporting any of the specific health problems; for example, women were more likely to report thyroid problems (12.1% vs. 6.2%, p < 0.00001) or arthritis/osteoporosis (32.9% vs. 15.5, p < 0.00001). Men, on the other hand, were more likely to report kidney/bladder problems (14.2% vs. 9.8%, p < 0.00001) or more hearing loss (14.9% vs. 9.9%, p < 0.00001). These patterns probably reflect a complex interaction of gender-related cancer diagnoses and co morbidities.
|Health problem||No. of patients||Women||Men||p-value|
|No.||% of all women||No.||% of all men|
Table IV depicts the 4 most frequently listed cancer types relative to specific reported health problems (we included cancer types that represented at least 10% of the total cohort, thus excluding “other,” multiple myeloma, neuroendocrine, unknown primaries and central nervous system tumors, which were the most infrequent). Arthritis/osteoporosis and heart problems were among the most frequently noted health problems. The pattern of perceived and reported health problems was different among the cancer types. For example, survivors of Hodgkin's disease prominently reported thyroid and lung problems (33.8% of responders with the diagnosis), whereas prior diagnosis of lymphoma was associated with frequent mention of memory loss (14.7%).
|Cancer type||Age at survey1||Health effect (% of patients with cancer type)|
|Most frequent||Second most frequent||Third most frequent||Fourth most frequent|
|Genitourinary||69.6 ± 13.4||Urinary (22.2)||Arthritis/osteoporosis (18.2)||Heart (16.5)||Cataracts (15.4)|
|Breast||66.3 ± 11.4||Arthritis/osteoporosis (37.6)||Cataracts (19.1)||Heart (15.3)||Thyroid (11.1)|
|Sarcoma||60.6 ± 14.0||Arthritis/osteoporosis (23.6)||Heart (18.1)||Circulation (12.4)||Hearing loss (10.0)|
|Gynecologic||65.4 ± 13.6||Arthritis/osteoporosis (31.1)||Urinary (18.5)||Cataracts (16.3)||Heart (13.3)|
|Acute leukemia||55.0 ± 13.7||Arthritis/osteoporosis (19.1)||Memory loss (14.7)||Migraine headaches (10.3)||Heart (10.3)|
|Melanoma||64.1 ± 12.7||Cataracts (13.6)||Heart (12.4)||Arthritis/osteoporosis (10.6)||Hearing loss (10.1)|
|Hodgkin's disease||50.9 ± 11.8||Thyroid (33.8)||Heart (26.3)||Arthritis/osteoporosis (14.4)||Lung (10.0)|
|Chronic leukemia||63.1 ± 11.7||Frequent infections (15.1)||Cataracts (17.8)||Arthritis/osteoporosis (17.8)||Heart (15.1)|
|Colorectal||71.2 ± 11.4||Arthritis/osteoporosis (21.5)||Cataracts (19.1)||Heart (15.4)||Hearing loss (13.8)|
|Head and neck||69.0 ± 12.0||Arthritis/osteoporosis (21.3)||Cataracts (19.1)||Thyroid (16.7)||Heart (14.0)|
|Lymphoma||63.6 ± 13.3||Arthritis/osteoporosis (25.3)||Cataracts (21.1)||Heart (17.7)||Memory loss (14.7)|
|Thyroid2||62.8 ± 13.5||Arthritis/osteoporosis (26.4)||Cataracts (13.2)||Heart (12.6)||Hearing loss (9.3)|
|Gastrointestinal||70.2 ± 10.5||Heart (18.6)||Cataracts (16.9)||Arthritis/osteoporosis (16.9)||Memory loss (15.3)|
|Lung3||68.0 ± 9.1||Arthritis/osteoporosis (24.4)||Heart (18.3)||Cataracts (18.3)||Hearing loss (17.1)|
The association between the time elapsed since the cancer was diagnosed and the reporting of specific health effects is depicted in Table V. Correlation coefficients were calculated to determine if there was a relationship between time from diagnosis and health effects. This analysis showed significant correlations between how long ago the cancer was diagnosed and a number of specific health effects; most became more frequent over time except memory loss.
|Health problem||Time from diagnosis (yr)||p-value|
Comparison with NHIS
A large number of health conditions are reported by the National Health Interview Survey (NHIS) and categorized by age, ethnicity/race and gender. Six health conditions reported in the NHIS were comparable with health conditions that were addressed in our cancer survivor survey: arthritis symptoms, diabetes, migraine or severe headaches, heart disease, cataracts and hearing loss. The frequencies of these conditions were used to estimate the number of affected individuals/1,000 population in order to compare the occurrence of the different health problems reported by the cancer survivors responding to our survey with their occurrence in the general population relative to age and gender; these results are summarized in Table VI.
|Health condition||Survivors||<451 NHIS||p-value||Survivors||45–641 NHIS||p-value||Survivors||≥651 NHIS||p-value|
A number of differences were noted between the survivors and the NHIS population both relative to gender and age. The incidence of all 6 health conditions appeared higher for male and female cancer survivors who were younger than 45 years when they responded to the survey, the reporting of arthritis and cataracts being most notable. With increasing age, the pattern of reported health conditions changed. However, for reasons not identified at present, the reported incidence of cataracts remained much higher among cancer survivors in all age groups and for both men and women.
As the population of cancer survivors continues to rise, the health profile and medical needs of this group will become increasingly relevant to our public health policies and priorities. Although data are accumulating regarding the psychosocial sequelae of cancer and cancer treatments and the quality of life of cancer survivors, there is limited information regarding lasting medical problems in adults with a history of childhood cancer5, 6, 7, 8 and even fewer data regarding lasting medical effects in survivors of adult cancers.9, 10 Furthermore, available reports tend to focus on relatively narrow aspects of late effects in specific cancer types or they tend to provide relatively short-term follow-up. There is a need to develop a comprehensive understanding of the scope and clinical relevance of perceived and significant differences or similarities between the population of cancer survivors and the population at large. In addition, there is a need to delineate the influence, if any, of a number of variables such as disease type, treatment type(s), gender, age, ethnicity and associated comorbidities, to name a few. In the present report, we have described the development of a health profile survey as 1 approach to developing such a body of knowledge. We have also presented the broadly demographic and clinical characteristics of a large group of adult cancer survivors who responded to this survey by mail.
Several observations are noteworthy, we believe. The response rate to the mailed survey (50.8%) was reassuringly robust and possibly an underestimate (we could not confirm, for example, that all mailed surveys were received), especially in view of the fact that these individuals are generally well persons who have infrequent contact with the cancer center and no particular motivation to answer a relatively long list of questions. In addition, almost one-third of the survivors provided us with additional contact information (address, telephone numbers), and many included often voluminous unsolicited comments about their cancer, their health and their opinions (data under analysis, not reported here). This indicates that cancer survivors are willing to provide information about their health and welcome this line of inquiry This response is in keeping with other investigators' findings. Pakilit et al.22 proposed using tumor registries for cancer survivorship research and showed that 5–10-year breast cancer survivors completed surveys about their health.
It was also interesting and reassuring to note that most survivors did not perceive overwhelming cancer-related health problems. Only one-third of respondents indicated that they believed cancer had affected their overall health. This pattern was somewhat different from the finding by Hewitt et al.23 that only 34% of survivors believed they were in good or excellent health. We were also reassured that, despite reports indicating frequent and significant depression in cancer patients,24, 25 self-reporting of psychological problems was uncommon in our study population. Variability in the populations of survivors who may comment about their health will probably create differences in the information derived.
Several patterns emerged from our analysis of the survey responses. We found that gender, age, length of time since diagnosis and cancer type affect the frequency and type of health effects experienced by long-term cancer survivors. Other investigators have noted significant late morbidity in Hodgkin's disease survivors,26 as well as correlations between age and adverse effects.27 We also found a high preponderance of health effects in Hodgkin's disease survivors; despite their young age, they frequently reported heart problems (26.3%), thyroid problems (33.8%), arthritis/osteoporosis problems (14.4%) and lung problems (10%).
Our data indicate that some cancer-related health effects (such as memory loss) become less prominent over time whereas others (such as hearing loss or cataracts) become much more prominent. Such changes may relate in part to the likelihood that very long-term cancer survivors are also older and subject to the health consequences of aging. Indeed, a limited comparison between the survivors and the general population indicated that the apparent incidence of several common health conditions varied with age at the time of the survey. Of note, the cancer survivors younger than 45 years reported a higher incidence of all health conditions; as age increased, the differences appeared to lessen, although reporting of cataracts remained prominent. Comparing survey response rates with information from the NHIS and other public health sources must be interpreted cautiously but can provide important insights about the health profile of cancer survivors. The survivors were most likely to report cancer-related health problems 5–10 years after diagnosis (37%), although 26% of the patients who were diagnosed >35 years earlier also reported problems.
The late health effects seen in cancer survivors may to some extent be due to inherent susceptibilities to other health conditions. However, it is quite possible that they are also related to the specific therapeutic regimens the patient underwent for treatment of the malignancy.10, 28 Specific organ system toxicities have been associated with disease, age and treatment characteristics and have included cardiopulmonary,29, 30, 31, 32 vascular,33 skeletal34, 35 and gonad abnormalities.36, 37 In the present report, we have not detailed correlations between specific therapies and medical health effects. However, the pattern of health effects seen in different diseases, such as hearing loss in relatively young testicular cancer patients or endocrine problems in Hodgkin's disease patients (data not shown), suggests that specific treatments for individual diseases may involve particular health effects.
Patients are increasingly living for many years after the diagnosis of cancer. An awareness of their history of cancer and cancer treatment often necessitates modification of generally held medical practice standards. Accordingly, there has been a paradigm shift away from cancer as an acute or terminal event toward cancer as a chronic condition. Cancer survivors are becoming much more prevalent in the population, but their unique medical health care needs are not well defined. Hollen et al.38 have proposed the creation of comprehensive specialty follow-up clinics for long term-cancer survivors. We agree wholeheartedly and have established such a program (the LACC medical clinic) at M. D. Anderson in response to this need. However, we cannot yet design appropriate medical follow-up guidelines for survivors of specific cancers because we do not have sufficient knowledge regarding the long-term health profiles for the different diseases and treatments. We have described 1 approach for the development of such a body of knowledge and have provided several broad characteristics of the health profile of cancer survivors. Much additional work is needed to develop accurate and comprehensive information about the lasting medical and social implications of the different types of cancer and cancer therapies. Information gathered from the survey presented in the present report can serve as a basis for additional, more focused inquiries and analyses of selected parameters such as disease type, treatment modality, and gender or ethnicity characteristics.
We thank all the patients who responded to the survey, their health care teams who provided excellent care, Sarah Taylor for contribution to the mailed survey, Dr. Carl DeMoore for valuable assistance in the study design and Roseann Martinez for manuscript preparation.
- 2Cancer Facts & Figures 2002. American Cancer Society, Inc. SEER Cancer Statistics Review, 1973–1998.
- 3Late effects of childhood cancer and its treatment. In: PizzoPA, PoplackDG, editors. Principles and practice of pediatric oncology, 2nd ed. Philadelphia: JB Lippincott, 1993. 1091–114., , .
- 21Summary health statistics for U.S. adults: National Health Interview Survey, 1997. National Center for Health Statistics. Vital Health Stat 2002; 10., , .
- 35Preventing bone loss and fractures. Coping 1998; 12: 68–9., .
- 37Adverse effects of treatment: gonadal dysfunction. In: DeVitaVTJr, HellmasS, RosenbergSA, editors. Cancer: principles and practice of oncology, 6th ed. Philadelphia: JB Lippincott; 2001. 2923–39., , .